Browsing by Author "Dr. Al Riordan, Committee Member"

Now showing 1 - 4 of 4

An Analysis of Hurricane Debby (2000) and the Impact of Vertical Shear on the GFDL Forecast Performance.
(2002-09-04) Rhome, Jamie Robert; Dr. Al Riordan, Committee Member; Dr. Sethu Raman, Committee Chair; Dr. Gary Lackmann, Committee Member
The need for improved tropical cyclone (TC) intensity guidance has never been greater given recent upward population trends in coastal areas. The difficulty in forecasting rapid intensity change remains one of the more challenging aspects of TC forecasting and was recently highlighted by the unexpected weakening of Hurricane Debby (2000) along the northern coast of Hispaniola on August 23, 2000. To address the need for improved understanding of rapid intensity change and the ability of dynamical TC models to accurately forecast intensity, a three-dimensional dynamical TC model (GFDL) is analyzed during the lifecycle of Debby. This was accomplished by first performing a comprehensive observational analysis making use of observed in-situ data as well as remotely sensed satellite data and derived products. Results from this analysis indicate that rapidly increasing vertical shear was the primary catalyst for the sudden weakening. Accordingly, vertical shear was analyzed within several operational simulations of the GFDL model near the time of weakening. This was accomplished by comparing the GFDL initial and forecast intensity with the National Hurricane Center official best track data as well as comparing the GFDL vertical shear with the AVN global analysis. Deviations in the GFDL intensity and vertical shear from the analysis data were considered to represent forecast error. These errors were then analyzed further by comparing the GFDL model forecast environmental wind field with a suite of observed data including GOES-8 satellite-derived winds, NOAA G-IV dropwindsondes, and upper-air observations supplemented by the GFDL initial analysis (F00). Results indicate that errors in vertical shear were nearly coincident with deviations in observed intensity versus forecast intensity. These deviations were primarily the result of a misrepresentation of the upper-level flow in the model due to an overdeveloped downstream upper-level ridge. Additionally, an erroneous anticyclone developed over the model storm in several cases, resulting in significant weakening of the upper level westerly flow and associated vertical shear. In this case, the downstream anticyclone was more intense and closer to the storm in nearly all simulations analyzed. These findings are similar to previous studies where a storm to environment interaction has been identified as the result of redistribution of latent heat release due to convection and the downstream advection of the associated low Potential Vorticity (PV) outflow. The misrepresentation of convection and the associated effects on the surrounding environment is identified as the primary factor for both track and intensity forecast errors by the GFDL model during Debby.
Evaluation of MM5 Forecasts of Near-Surface Parameters: Sensitivity to Land-Surface Parameterization and Planetary Boundary Layer Schemes
(2002-09-30) Cerjak, Jason Robert; Dr. Al Riordan, Committee Member; Dr. Jerry Davis, Committee Co-Chair; Dr. Gary Lackmann, Committee Co-Chair
The specific purpose of the research is to evaluate the performance of the MM5 model in the forecasting of near-surface parameters, such as 2-meter temperature, 2-meter dew point, and 1000-850 mb thickness. The evaluation will include a comparison of the MM5 against the Eta model, and a comparison of the forecasting skill of the MM5 with three different land-surface parameterization schemes. Three different soil moisture scaling techniques will be applied to the MM5, and their forecasts will be evaluated against observations taken from 7-9 December 2001. The MM5 displayed an inability to capture the full magnitude of the diurnal cycle of 2-meter temperature and dew point. The Eta model performed better than the MM5 in the forecasts of near-surface parameters. The MM5 forecasts of near-surface parameters can be improved by adjusting the vertical profile of the soil moisture in the model initial conditions. By removing soil moisture from the initial conditions of the MM5, a more realistic Bowen ratio was simulated, leading to an improved forecast of the diurnal cycle of temperature and dew point. The results of this research suggest that the cause of the damped diurnal cycle in the MM5 forecasts may be inadequate ventilation of the upper planetary boundary layer, an inaccurate representation of surface evapotranspiration, or incorrect assignment of soil type and land use categories.
A Statistical and Synoptic Investigation of Tropical Cyclone Intensity Changes Over the Gulf Stream Along the Southeast Coast of the United States
(2003-10-06) Bright, Robert James Jr.; Dr. Lian Xie, Committee Chair; Dr. Len Pietrafesa, Committee Member; Dr. Al Riordan, Committee Member
Cases of unexpected rapid intensity changes of tropical cyclones (TCs) close to the coast of the United States (U.S.) such as Opal (1995), Bertha (1996), and Lili (2002) have prompted more research into the factors that affect TC strength. It is well known that along with large-scale environmental and storm-scale internal influences, TC intensity is also influenced by interactions with the underlying ocean. In particular, warm ocean currents such as the Loop Current (LC) in the Gulf of Mexico and the Gulf Stream (GS) along the U.S. East Coast can provide these storms with extra energy through enhanced heat and moisture fluxes. However, the extent to which each of these factors is important is not as well understood. Although previous studies have investigated the potential impact of the LC and its associated warm-core eddy in helping to strengthen TCs, none have focused exclusively on what impact the GS may have. Thus, this research attempts to document the importance of the GS in influencing the strength of TCs approaching the Southeast Coast of the U.S. This was accomplished by performing a climatological study of 48 TCs that crossed the GS during the period 1944-2000 as well as case studies of three recent landfalling hurricanes in North Carolina. In addition, an empirical-statistical prediction scheme was developed to forecast the net intensity changes of TCs over the GS. The statistical analysis based on historical "best track" data suggested that the GS helped to either enhance or at least maintain the strength of most storms. In fact, 75% and 72% of TCs either intensified or maintained their strength over the GS according to maximum wind speed (MWS) and minimum central pressure (MCP), respectively. Moreover, most of these storms experienced greater intensification rates than they had prior to reaching the GS. Composite analysis revealed that strengthening of TCs over the GS was typically associated with a trough northwest of the TC similar to the results of previous studies. It is believed that the trough could contribute positively to the storm by steering it along the GS, which would allow for increased TC-GS interaction, and/or through direct TC-trough interaction. On the other hand, weakening was associated with two different synoptic scenarios. The first is a situation in which no trough is present northwest of the TC, thus allowing the storm to move across the GS rather than along it. This would essentially limit the TC-GS interaction. In fact, the climatological study revealed that the weakening cases spent less than half the time over the GS on average than the storms that intensified. The second scenario is one in which the trough northwest of the TC is stronger than for the intensifying cases. Such a setup would likely lead to increased wind shear over the storm. Thus, the negative effect of shear would likely dominate any positive contributions from the trough (e.g., increased upper-level outflow, enhanced momentum fluxes, etc.). Based on the results of the climatological study, a regional TC intensity prediction scheme was developed to forecast the net intensity changes of TCs over the GS. Stepwise multiple linear regressions were performed to fit climatological, persistence, and synoptic predictors to the observed net MWS and MCP changes, respectively. Results indicate that the storm's duration over the GS, initial intensity, and previous 12-h MWS change were the statistically significant parameters in the MWS model. The important variables in the MCP model include the storm's duration over the GS, 12-h MCP change, and the magnitude of the 850-200-hPa shear at the time the storm left the GS. The coefficients of the parameters in each model are physically consistent and agree with the results of the climatological study. For example, the longer a TC spends over the GS the more likely it is to intensify. Also, intensification is more likely if the storm is already intensifying prior to the GS. Forecasts based on the MWS model were then compared to average forecast errors from various operational models as well as the National Hurricane Center official forecasts. Overall, the MWS model performed well despite it being a simple persistence-type model. Case studies of three recent landfalling hurricanes in North Carolina were also performed to highlight the different types of intensity changes experienced by TCs over the GS. Hurricane Bertha (1996) rapidly intensified prior to landfall as it crossed the GS and interacted with a short-wave trough. It appeared this event was due to a combination of enhanced TC-GS interaction while in a favorable upper-level environment. Hurricane Bonnie (1998) basically maintained its strength over the GS. However, the storm did experience a slight MCP decrease after the GS, but its proximity to the coast likely limited any chance for intensification. On the other hand, Floyd weakened while traversing the GS, albeit more slowly than it had been prior to reaching the GS. Dry air entrainment and increased vertical wind shear seemed to be the likely causes for Floyd's general weakening trend during that period. However, the storm then maintained its MWS after crossing the GS prior to making landfall. Overall, the results of the three case studies support the findings of the climatological study and other previous studies that indicate the positive influence warm ocean features can have on TC intensity.
Three-Dimensional Radar and Total Lightning Characteristics of Mesoscale Convective Systems
(2003-08-11) McCormick, Tracy Lynn; Dr. Al Riordan, Committee Member; Dr. Gary Lackmann, Committee Member; Dr. Lawrence D. Carey, Committee Chair
The radar and electrical characteristics of three linear leading convective/trailing stratiform midlatitude mesoscale convective systems (MCSs) that passed through Dallas-Fort Worth, Texas on the following dates are examined: 1) 7-8 April 2002, 2) 12-13 October 2001, and 3) 16 June 2002. Quantitative results from the April and June MCSs are presented, but data problems with the October MCS restricted partitioned analysis to qualitative results. The convective line produced ~69% and ~93% of the total cloud-to-ground (CG) lightning flashes in the April and June MCSs, respectively. The convective line CG flash rate averaged 12.3 flashes min-1 (53.6 flashes min-1) in the April (June) case study, and only 7.5% (2%) of these flashes were positive in polarity. Lightning Detection and Ranging (LDAR II) source data identified two main electrically-active regions present within the convective line in the following temperature layers: 1) 0 to -25 °C, and 2) -35 to -55 °C. The lower region (1) was most likely a combination of the main negative and the lower positive charge centers of the thunderstorm tripole, and the upper region (2) was most likely the upper positive charge center of this tripole. Convective echo volume aloft (≥ 30 dBZ, 0 to -40 °C) was strongly correlated to convective lightning activity, suggesting that the presence of strong updrafts and differential sedimentation caused convective line electrification via the non-inductive charging (NIC) mechanism. The stratiform region CG flash rate averaged 2.2 flashes min-1 (4.5 flashes min-1) in the April (June) case study, and ~45% (~27%) of these flashes were positive in polarity. LDAR II source data identified one primary electrically-active layer (at -10 to -25 °C) that was sloped from the upper portions of the convective line rearward to just above the bright band in the stratiform region. A small and spatially distinct secondary electrically-active layer (at ~ -40 °C) was located towards the rear of the stratiform region. These two layers had smaller average source concentrations than the convective line had, resulting in significantly less lightning production in the stratiform region than in the convective line. Hydrometeor trajectory analyses using storm-relative vertical and horizontal motions determined from synthetic dual-Doppler results indicate that the these two stratiform region layers become electrified by a combination of 1) positive charge advection from the upper-positive convective charge center to the stratiform region and 2) stratiform in situ charging via NIC, likely creating an inverted dipole. This inverted dipole may explain why the +CG flash percentage was significantly higher in the stratiform region than in the convective region (where a normal dipole is present). In addition, stratiform echo volume aloft (≥ 25 dBZ, -10 to -40 °C) was strongly correlated to stratiform lightning activity, suggesting that differential sedimentation as a result of the presence of larger ice aggregates (i.e. Z > 25 dBZ) at these temperatures was required for stratiform region electrification via both charge advection and in-situ charging.